180 PEINGIPLKS OP PLANT-TBEATOLOGY. 



nearly allied as they are to the Amaryllidacese, not only 

 has the corolla, but the calyx also, been derived from 

 the andrcBcium. 



Other Theories of the Corona. 



The earlier views as to the nature of the corona of 

 Narcissus were as follows. Masters regarded it as 

 composed of the first two whorls of petaloid stamens 

 fused together. Celakovsky comments on this as being 

 the least probable view of all, though never definitely 

 refuted ; but certainly never proved. Baillon, relying 

 exclusively on developmental data, regarded it as a 

 cupular outgrowth of the receptacle, of the nature of 

 a disk. But here, as ever, the developmental method 

 is altogether misleading, its followers being ignorant 

 of congenital fusion-phenomena such as this case 

 affords. Doll, W. Gr. Smith, and Bichler regarded the 

 corona as a ligular outgrowth of the perianth com- 

 parable to the corona of Caryophyllaceae ; yet Pax 

 regarded the very similar corona of Pancratium, a 

 member of the same order, as composed of the fused 

 stipules of the stamens. As Celakovsky points out, 

 the systematic-comparative, like the developmental 

 method, can throw no real light on the problem. This 

 can only be done by means of a study of the teratological 

 phenomena. 



The Abnormal "Corona'^ and Anther-Structure. 



Some light may probably also be thrown on the 

 morphology of anther- structure by a contemplation of 

 some of these structures exhibited by the petaloid 

 stamens of Narcissus. In some the corona-like ventral 

 excrescence becomes invaginated towards the upper sur- 

 face of the lamina, thus reminding us of a similar basal 

 invagination in the saxifrage-leaves above described, 

 and, above all, in the normal leaves of several species 

 of Galtha. Also the pocket-shaped forms of stamen 

 (PI. XLIV, fig. 6; and fig. 132, on left) are exactly 

 equivalent to the similarly-shaped leaves of Saxifraga 



